OUP user menu

Construction of bioluminescent cyanobacterial reporter strains for detection of nickel, cobalt and zinc

Loredana Peca, Péter B. Kós, Zoltan Máté, Andrea Farsang, Imre Vass
DOI: http://dx.doi.org/10.1111/j.1574-6968.2008.01393.x 258-264 First published online: 1 December 2008


Two whole-cell bioluminescent reporters were constructed by fusing the reporter genes luxAB with the Co2+ and Zn2+ inducible coaT promoter or the Ni2+-inducible nrsBACD promoter, respectively, in the genome of Synechocystis sp. PCC 6803. The obtained reporters, designated coaLux and nrsLux, respectively, responded quantitatively to metal ions. After 3 h incubation at 40 μmol m−2 s−1 visible light, the detection range of coaLux was 0.3–6 μM for Co2+ and 1–3 μM for Zn2+. Incubation in darkness increased the detection range by about four times. The nrsLux reporter was specific to Ni2+, with a detection range of 0.2–6 μM. However, its activity was inhibited by Zn2+ with a half maximal inhibitory concentration c. 6 μM, and totally inhibited by darkness. This is the first whole-cell Ni2+-specific reporter with a clear dose–signal relationship. In a soil-like mixture of different chemical and oil industry wastes, the coaLux reporter strain detected about 90% of the zinc content of the sample. This study demonstrates the potential for development of a rapid, simple and economical field assay for nickel, cobalt and zinc detection using the coaLux and nrsLux reporters.

  • cyanobacteria
  • Synechocystis PCC 6803
  • bioluminescent bioreporter
  • cobalt
  • nickel
  • zinc


Cyanobacteria are photosynthetic prokaryotes that require a variety of metal cations such as Cu2+, Ni2+, Fe2+ and Zn2+ to maintain their cellular metabolism (Cavet et al., 2003; Baptista & Vasconcelos, 2006). Because all metal species, including the essential ones, are toxic at high concentrations, sensitive regulation of metal uptake, storage, allocation and detoxification is required to maintain their homeostasis. Cyanobacterial cells have complex metal resistance mechanisms that include enzymatic detoxification, sequestration and exclusion via active transport with either CPx-ATPases or chemiosmotic efflux systems (Silver & Phung, 1996; Argüello et al., 2007).

The widely used model organism Synechocystis sp. PCC 6803 (henceforth referred to as Synechocystis) possesses a chromosomal gene cluster involved in Ni2+, Co2+ and Zn2+ tolerance. The divergently transcribed coaT and coaR genes (Fig. 1) mediate inducible resistance to Co2+. We have previously shown that coaT was similarly induced by both Co2+ and Zn2+ following a short 15-min incubation (Peca et al., 2007). The coaT-disrupted mutant showed a decreased resistance to Co2+, but not to Zn2+ (Garcia-Dominguez et al., 2000) and an increased accumulation of 57Co in the cytoplasm (Rutherford et al., 1999). Synechocystis double mutants ΔziaΔcoa displayed similar sensitivity to Zn2+ as Δzia single mutants, suggesting that there is no residual transport of Zn2+ by CoaT (Borrelly et al., 2004). These lines of evidence, together with the clear homology of the coaT product with cation-transporting P-type ATPases suggested that CoaT is a Co2+ efflux P-type ATPase. It has also been demonstrated that coaR encodes a transcriptional activator of coaT whose carboxyl-terminal Cys–His–Cys motif is required for cobalt sensing (Rutherford et al., 1999).

Figure 1

Schematic representation of the cloning strategy used for construction of nrsLux and coaLux reporters. The upper part of the figure shows the structure of the chromosomal nrs and coa operons and the sequences of PCR primers (shown in bold) used to amplify the nrsBACD and coaT promoters, along with their regulator gene sequences. Arrows above the ORFs indicate the directions of transcription. The nonhybridizing bases of the primers are underlined and SalI recognition sites are boxed. The lower part of the figure shows the plasmid map of pND6luxAB. HS1 and HS2 represent the flanking sequences required for double homologous recombination. The HS1 sequence spans from 155889 to 156981 and HS2 sequence spans from 156978 to 157478 in the Synechocystis genome according to conventional numbering in the database CyanoBase. Cloning of SalI-digested PCR-amplified sequences into the SalI/SmaI-digested pND6luxAB allowed the construction of nrs–luxAB and coa–luxAB transcriptional fusions. Amp and cam refer to genes conferring resistance to ampicillin and chloramphenicol, respectively.

Adjacent to coa genes there are two divergent operons, nrsRS and nrsBACD (Fig. 1), coding for proteins involved in nickel resistance. The nrsBACD genes are induced in the presence of Ni2+ and Co2+ (Garcia-Dominguez et al., 2000; Lopez-Maury et al., 2002). However, their function has not been fully elucidated. Reduced Ni2+ tolerance was observed following disruption of nrsA and nrsD. It has been hypothesized that nrsA and nrsB products form a Ni2+ efflux system, based on their homology with Ralstonia eutrophus czcA and czcB protein products (Nies et al., 1989). NrsD is a putative member of the major facilitator superfamily of permeases involved in Ni2+ export, whereas the protein product of nrsC shows no homology to proteins encoded by the czc or related operons. The nrsR and nrsS ORFs seem to form a two-component system controlling the nickel-dependent expression of the nrsBACD operon. The nrsRS disruption decreases Ni2+ tolerance and abolishes the Ni2+-dependent inducibility of the nrsBACD operon (Lopez-Maury et al., 2002).

In the last two decades, bacterial resistance mechanisms against various metal and metalloid cations have been used to construct living whole-cell biosensors or bioreporters (for the principle, see Virta et al., 1998; for an extensive review, see Harms, 2007). Using photoautotrophic cyanobacteria for whole-cell biosensing purposes represent an advantage over the use of heterotrophic microorganisms because they can grow on low-cost media and require little maintenance (Erbe et al., 1996; Bachmann, 2003). Because Synechocystis is easy to transform and its full genomic sequence is available (Kaneko et al., 1996), it represents a suitable organism for whole-cell bioreporter construction. Consequently, iron and nitrogen sensor strains were previously constructed using lux genes in this cyanobacterium (Kunert et al., 2000; Mbeunkui et al., 2002).

In this study in order to develop Co2+, Zn2+ and Ni2+ sensors, the coaT and nrsBACD promoters, together with the coaR and nrsR regulatory gene sequences, respectively, were fused to promoterless luciferase (luxAB) reporter genes. These fusions were introduced into the genome of Synechocystis constitutively expressing luxCDE genes coding for the fatty acid reductase complex that produces the luciferase aldehyde substrate. The response of the obtained bioreporter strains coaLux and nrsLux to a range of various metal and metalloid salts was investigated, and their suitability as whole cell biosensors is discussed.

Materials and methods

Growth conditions and metal salt treatment

Synechocystis cells were grown photoautotrophically at 40 μmol m−2 s−1 and 30 °C in BG-11 medium (Rippka et al., 1979) in a 3% CO2-enriched atmosphere. The starter culture was prepared by inoculation of 1.5 mL frozen stock culture into 50 mL BG-11 liquid medium containing 25 μg mL−1 spectinomycin and 5 μg mL−1 chloramphenicol. From the fully grown starter culture, 1 mL was inoculated into 200 mL BG-11 without antibiotics and incubated until mid-log phase (OD720 nmc. 0.6). Metal and metalloid salt treatments were carried out in microtiter plates in BG-11 supplemented with ZnSO4, CdCl2, NiCl2, CoCl2, NaAsO2, KH2AsO4, CuSO4, Cr2(SO4)3 or Na2CrO4. No glucose was added to the medium.

Construction of coaLux and nrsLux reporter strains

A sequence from upstream of the coaT gene containing the coaT operator–promoter and coaR regulator gene was amplified from the genomic DNA using the primers displayed in Fig. 1, by introducing a SalI recognition site at the 5′ end of the amplicon. Similarly, the upstream region of the nrsB gene, containing the nrsB operator–promoter and the sequence of the nrsR regulator gene (Fig. 1) was amplified. The amplicons were ligated upstream of the promoterless luciferase genes luxAB from Vibrio harveyi, into the SalI/SmaI digested vector pND6luxAB (Fig. 1). The construct was used for transformation of the ‘substrate’ strain, a Synechocystis strain harboring the fatty acid reductase complex genes luxCDE from V. harveyi, necessary for the production of the luciferase aldehyde substrate. The constructs were integrated into the Synechocystis genome along with a chloramphenicol cassette via homologous recombination at a neutral site (Aoki et al., 1995), using the corresponding sequences HS1–HS2 in the pND6luxAB integration vector (Fig. 1). The growth rates in BG-11 of the obtained coaLux and nrsLux reporter strains were comparable with that of the wild type (data not shown). The pND6luxAB vector and the ‘substrate’ strain were kindly provided by M. Kis (unpublished data). The stoichiometry of the regulatory protein vs. the regulated promoter copies necessary for ideal functioning of the regulator is not known. In order to not alter the original 1 : 1 ratio of promoter/regulator gene sequence, the regulatory gene sequences were cloned along with coaT and nrsBACD promoters. The strains and vectors constructed hereby are freely available to the scientific community upon request.

Bioluminescence assay

Metal salt treatments were carried out in 96-well black microtiter plates with low autofluorescence (Packard HTRF) in a 300-μL-per-well volume (200 μL culture and 100 μL salt per well of microtiter plate). The plates were covered with needle-punctured transparent foil and incubated for 3 h in light (40 μmol m−2 s−1) or darkness (in a CO2 permeable opaque box) in a 3% CO2-enriched atmosphere. Assays were performed with four parallel samples at 25 °C. Luminescence intensity was determined with a Top Count NXT luminometer (Packard Instruments) and is expressed as counts per second. The relative luminescence induction was calculated by dividing the mean luminescence signal of a treated sample by the mean luminescence signal of the untreated sample.

Soil sampling and preparation

The soil-like material used for this study consists of a mixture of different chemical and oil industry wastes from Almásfüzitő bauxite residue disposal area in NE Hungary. Its heavy metal content considerably exceeds the threshold limits allowed for soils by Hungarian environmental regulations. Samples collected in triplicate from the composting piles were dried, grinded, the coarse material was removed and the remaining material was passed through a 2-mm sieve. To assess the exchangeable, acid-soluble fractions of Ni2+, Co2+ and Zn2+, one-step acetic acid extraction of the material samples was carried out as follows (Bódog et al., 1996). Aliquots of 1.0 g soil were mixed with 40 mL 0.11 mol L−1 CH3COOH. The suspension was shaken for 16 h at room temperature and then centrifuged for 15 min at 1500 g. The supernatant was preserved by the addition of 0.04 mL HNO3 and made up to 50 mL. The metal content was determined using an atomic absorption spectrophotometer (AAS) (Perkin-Elmer model 3110). Each soil sample was analyzed in triplicate. For the bioluminescence measurements, the pH of the CH3COOH-extracted sample was adjusted to 7.5 with 1 M Tris-HCl pH 8.0.


Specificity of the bioreporter strains

Luminescence measurements were optimized to detect the presence of metal ions with a simple protocol. Mid-log phase Synechocystis coaLux and nrsLux cultures were incubated in light with the salts listed in Materials and methods. We found that the optimal induction time was 3 h. Longer incubation time did not significantly increase the relative induction. Specificity of the bioluminescent response was investigated in BG-11 medium containing different salts for which the minimal (ICmin) and maximal inhibitory concentrations (ICmax) were previously established (Peca et al., 2007). Two concentrations in these ranges where chosen as indicated in Fig. 2. From the tested compounds, only Zn2+ and Co2+ induced a significant bioluminescent response in the coaLux strain (Fig. 2a), whereas the response of the nrsLux strain was specific to Ni2+ (Fig. 2b).

Figure 2

Specificity of the bioluminescent response of the coaLux (a) and nrsLux (b) bioreporter strains to different heavy metal and metalloid salts. Cells were incubated in light for 3 h in BG-11 medium supplemented with known concentrations of selected heavy metals and metalloids. Assays were performed with four parallel samples. The SEs are indicated. CPS, counts per second.

Sensitivity of the bioreporter strains

Both the coaLux and nrsLux strains showed dose-dependent responses to the metal salts added into the culture medium. When incubated in light (40 μmol m−2 s−1), the coaLux sensor responded to Co2+ and Zn2+ with a detection range of c. 0.3–6 μM (Fig. 3b) and c. 1–3 μM (Fig. 3a), respectively. The luminescence increased gradually with increasing concentrations of Co2+ and Zn2+ up to 6.4 μM (relative luminescence induction of c. 70-fold) and 3.2 μM (relative luminescence induction of c. 25-fold), respectively, followed by a decrease to the background level at 15 μM. In mixed samples of Co2+ and Zn2+, the coaLux reporter strain responded in an additive manner (data not shown). We found that the absolute signal intensities correlated with the growth stage from early to late exponential phase and hence the cell numbers, but the pattern and relative induction of the luminescent signal was independent of the age of the culture (data not shown). The nrsLux sensor showed a detection range for Ni2+ of c. 0.2–6 μM. The maximum response was c. 50-fold induction of luminescence intensity at 6.4 μM Ni2+ in the medium (Fig. 3c). At higher Ni2+ concentrations, the luminescence signal gradually decreased, with 15 μM Ni2+ reducing the light emission to the background level. We observed a background luminescence emission for the coaLux reporter strain (two to three times higher than the wild-type luminescence emission), presumably due to the leakage of coaT promoter (data not shown).

Figure 3

The bioluminescent response of coaLux (a and b) and nrsLux (c) following light (open symbols) and dark (closed symbols) incubation. Cells were incubated for 3 h in BG-11 medium supplemented with Zn2+ (a), Co2+ (b) or Ni2+ salts (c), and the bioluminescence was measured as described in Materials and methods. Each point represents the mean of four parallels. The SEs are indicated.

The activity of the coaLux and nrsLux reporter strains was also evaluated in darkness. In the case of the coaLux reporter, the luminescence peaked at higher concentrations than in light-incubated samples and the peak position was stable over the 8-h range. Because of these changes in the shape of the luminescence response curves, the detection range increased by about four times: up to 26 μM Co2+ (relative luminescence induction of c. 25-fold) and up to about 13 μM Zn2+ (relative luminescence induction of c. 12-fold) (Fig. 3a,b). On the contrary, the nrsLux reporter strain showed no detectable luminescence when incubated in darkness with the tested concentrations of Ni2+ (Fig. 3c). This effect is most likely caused by a dramatic decrease of the nrsB promoter-driven expression of the lux genes as shown by the 10 × lower maximal luxA expression in darkness than in light according to quantitative reverse transcriptase-PCR assessment (data not shown).

Effect of Zn2+ on Ni2+-induced response in nrsLux

Because zinc and nickel pollution coexist in many sites, the nrsLux reporter strain in mixed samples was tested using Ni2+ and Zn2+ salt standards. Incubation was carried out for 3 h after the addition of metal salts. The Ni2+ concentration was maintained at 8 μM that produces a high luminescence signal in the nrsLux reporter strain while the Zn2+ concentration was varied from 0 to 16 μM, and the luminescence monitored. Interestingly, the Ni2+-specific response was markedly reduced in the presence of increasing Zn2+ concentrations from 2 μM, which produces about 20% inhibition, to 16 μM, which reduced the Ni2+-induced luminescence signal by about 90% (Fig. 4). The half maximal inhibitory concentration of Zn2+ (IC50) on the Ni2+-induced bioluminescence response was about 6 μM.

Figure 4

Concentration response curve for the inhibition by Zn2+ of Ni2+ luminescence response in the nrsLux reporter strain. Data are expressed as percentage of response in the presence of 8 μM Ni2+ alone. Each point represents the mean of data obtained in four wells for each concentration. Vertical lines show the SE. When not shown, the SE was smaller than the symbol. The solid line represents a formal fitting of the experimental data with an exponential function. The dotted lines indicate the IC50 value of Zn2+.

Quantification of bioavailable zinc in soil samples

In order to assess the performances of coaLux and nrsLux reporter strains, acetic acid extracts of contaminated soil, containing an average of 1.88 μM Ni2+, 0.72 μM Co2+ and 188.7 μM Zn2+, as determined by AAS, were used. Cells were incubated with 1.11-fold serial dilutions of the acetic acid extracts for 3 h in light, and then the bioluminescence was assayed. For calibration purposes, the reporter strain cells were incubated with 1.11-fold serial dilutions of a ZnSO4 standard solution. The response curve obtained with the standard samples could be fitted with a linear function with good correlation (R2=0.95) in the 1.6–2.5 μM range (Fig. 5). There is an inverse proportional relation between the dilution of an environmental sample and the bioluminescent signal for the linear ascending part of the sigmoidal curve. Calculations of Zn2+ concentration in the environmental sample dilutions were performed using the linear calibration curve. The results obtained by AAS and the luminescence-based detection technique showed a good correlation (Table 1). The concentration of Zn2+ calculated from the luminescence response of the coaLux reporter represented an average of 91.5 (±1.4 SD) % of the AAS-determined Zn2+ concentration. The Co2+ content of the environmental sample (0.72 μm) was not taken into account, because its value was negligible when compared with the Zn2+ concentration. The soil samples contained 1.88 μM Ni2+ as well, which is in the detection range of the nrsLux strain. However, no luminescence signal was detected upon incubation of nrsLux with the undiluted acetic acid extract of the contaminated soil, presumably due to the inhibitory high level of zinc content, exceeding by about 17 times the IC50 value.

Figure 5

Bioluminescence response of the coaLux reporter strain to a zinc-containing environmental sample extract. The reporter strain was incubated with 1.11-fold serial dilutions of ZnSO4 and acetic acid extract from a polluted soil sample for 3 h in light, and then the bioluminescence was assayed. Linear range of coaLux response to the standard sample dilutions and the equation of linear regression are presented. The extent of luminescence response to the different dilutions of the environmental sample extract is shown by the horizontal lines, whereas the Zn2+ concentrations calculated from the regression equation are shown by vertical lines (see the values in Table 1). The experiment was performed with four parallel luminescence assays. The SEs are indicated. CPS, counts per second.

View this table:
Table 1

Comparison of zinc content in the acetic acid-extractable fraction of an environmental sample as determined by AAS and by coaLux reporter strain

Fold dilutionZinc concentration (μM)
AASBiosensing (coaLux reporter)
  • Liquid suspensions of the reporter strain were incubated with different dilutions of the acetic acid extract of environmental sample, the bioluminescence induction was determined and the corresponding Zn2+ concentration was calculated from the equation of the standard curve displayed in Fig. 5.

  • * Fold dilution of the acetic acid-extractable fraction of an environmental sample with a content of 162 μM zinc, as determined by AAS.

  • Zinc concentration measured by AAS.

  • Zinc concentration calculated from the standard curve.


Since researchers in Sayler's lab obtained the first recombinant bioluminescent bacterial sensor for naphthalene (King et al., 1990), numerous whole-cell reporter organisms have been developed for determining a variety of metal cations in polluted soils or wastewater samples. A common characteristic of these bioreporters is their response to several related metal cations, and inhibition at high analyte concentrations that restricts their use to subtoxic concentrations (Nies et al., 1989). Several bioreporters that are able to sense Zn2+ and Co2+ besides other metal cations have been developed to date, and they rely on Synechococcus sp. smt (Huckle et al., 1993) and Escherichia coli znt (Riether et al., 2001) resistance systems.

The coaLux reporter strain presented in this study is based on the coa resistance system from Synechocystis, inducible only by Co2+ and Zn2+ from a range of tested metal and metalloid cations. Its detection range is about 0.3–6 μM for Co2+ and about 1–3 μM for Zn2+ when incubated in light. The shape of the concentration-dependent luminescence response for Zn2+ and the detection range are similar to those obtained with the cyanobacterial sensor based on Synechococcus smt-luxCDABE transcriptional fusion (Erbe et al., 1996). Incubation in darkness, in case of the coaLux reporter strain, has the advantage of extending the detection range up to 25.6 μM Co2+, and up to about 12.8 μM Zn2+.

The applicability of coaLux bioreporter for Zn2+ detection has been tested using polluted soil-like sample material collected from a composted mixture of different chemical and oil industry wastes. The concentration of Zn2+ detected by the coaLux reporter strain represents about 90% of the AAS-determined zinc concentration that can be interpreted as the bioavailable fraction. Having cloned the aldehyde synthase genes luxCDE into the host organism, we eliminated the need of external substrate addition, making the detection process easier and more convenient to handle.

The increasing utilization of nickel in modern technologies leads to an accumulation of nickel compounds in the environment, which may present a serious hazard to human health. Therefore, a cheap and effective detection method for environmental monitoring of nickel would be very useful. We created the nrsLux strain, whose selectivity for Ni2+ was demonstrated in a multiple metal assay.

The nrsLux reporter was inhibited by Zn2+ in a concentration-dependent manner, with an IC50 value of c. 6 μM. On the other hand, although coaT transcript is equally induced by Zn2+ and Co2+ (Peca et al., 2007), the bioluminescent response of coaLux strain to Zn2+ is much lower than that to Co2+ (see Fig. 3a and b). These observations together support the hypothesis that zinc ions interfere with bioluminescent reaction. It has been established that V. harveyi luciferase contains a highly reactive cysteinyl group at the aldehyde site and the structural integrity of its microenvironment is crucial to luciferase activity (Nicoli et al., 1974; Paquatte et al., 1988). It is known that cysteinyl residues are often involved in binding metal ions such as zinc. Therefore we hypothesize that zinc cations form nonspecific complexes with the active center cysteinyl residue of the luciferase, interfering with the aldehyde substrate binding and bioluminescence production.

The activity of the nrsLux construct was totally inhibited by darkness, probably caused by a significant decrease in luxA transcript level. The detection range from 0.2 to 8 μM Ni2+ matches the nickel concentration limit admitted in the drinking water specified by WHO's Guidelines for Drinking-Water Quality (0.07 mg L−1=1.19 μM). Zinc concentration in tap water is seldom above 0.15 μM, which is about 70 times less than the IC50 for Zn2+-induced inhibition of the nrsLux reporter. Therefore, the nrsLux reporter presented here, which is the first Ni2+-specific whole cell reporter with a clear dose–signal relationship, is potentially useful in the development of a biosensor for Ni2+ detection in drinking water.


We thank Mihály Kis (Biological Research Center, Szeged) for providing the ‘substrate’ strain and the pND6luxAB vector. We are grateful to Ferenc Nagy for providing access to a Top Count NXT luminometer. Funding for this project was provided by a grant from the Hungarian National Granting Agency OTKA (K68036).


  • Editor: Hermann Bothe


View Abstract